- munk-a parentCould you clarify what you mean by design questions? I do agree that GPT5 tends to have a better agentic dispatch style for math questions but I've found it has really struggled with data model design.
- As a developer - ChatGPT doesn't hold a candle compared to claude for coding related tasks and under performs for arbitrary format document parsing[1]. It still has value and can handle a lot of tasks that would amaze someone in 2020 - but it is simply falling behind and spending much more doing so.
1. It actually under performs Claude, Gemini and even some of the Grok models for accuracy with our use case of parsing PDFs and other rather arbitrarily formatted files.
- I think most people are aligned on AI being in a bubble right now with the disagreement being over which companies (if any) will weather the storm through the burst and come out profitable on the far side.
OpenAI, imo, is absolutely going to crash and burn - it has absolutely underwhelming revenue and model performance compared to others and has made astronomical expenditure commitments. It's very possible that a government bailout partially covers those debts but the chance of the company surviving the burst when it has dug such a deep hole seems slim to none.
I am genuinely surprised that generally fiscally conservative and grounded people like Jensen are still accepting any of that crash risk.
- You are correct that caring is important - but it also isn't your responsibility at the end of the day. If you don't care you're doing it wrong - if you let it eat you up inside whenever anything goes wrong you're also doing it wrong.
Work-life balance is mostly talked about in terms of time commitments but there is also an emotional commitment you need to balance. It's unhealthy to be too far in either extreme and, especially folks that are naturally empathetic, should be more wary of falling into the trap of overinvesting in a workplace and suffering mentally for it.
- > Try "I am unable to attend meeting without an agenda. Let me know when one has been posted." in your decline message.
If you have a good manager you can often CC them or quote them in your response as well "Sorry, I'm busy with project work and Sarah wants me to stay focused to hit our deadlines. If we're going to need to budget time outside of it I'll need a clear agenda to offer as a rationale to my stakeholders."
I think it really helps to sell this if you've got casual impromptu voice calls as a norm in the company. If it was really just a quick thing then throw up a hangout for us to chat - if it's worth scheduling a meeting for it's certainly worth actually putting together an agenda.
As an aside - my company recentlyish switched from google to ms for calendar management and (among many things MS is terrible at) the fact that agendas aren't immediately visible in meetings on your calendar is the absolute worst UX decision.
- Yup - they touch on proxy relationships where you have a few trusted reporters to break the crowd into cohorts that you can mentally simplify but whenever you do this you need to accept that it won't be complete. You should expect and make room for occasional noise from the fifty people behind your one trusted reporter because the problem could always lie with the reporter themselves.
- I wish LinkedIn was that good. In my eyes LinkedIn has become "Facebook but with resumes" - I accept that I should have a profile on there for resume visibility but there are so many features built into that site that just should not exist and so many genuinely valuable things around job seeking they could do that they're simply not.
It's an excellent example of a product that could be massively improved by just removing things - look back to early linkedIn days where their email notifications actually meant there was likely something relevant that you care about there. Now they've created a platform where the valuable is buried in piles of irrelevant slop.
- I think it's more just a bizarre platform for us to gawk at. I'm not really certain why LinkedIn even has social features available if it's purporting to be a professional space and especially a professional space that is going to be your first impression for a new job prospect. Maybe having a loud profile is a positive to some sorts of recruiters but posting anything beside resume information on that site just seems like a guaranteed malus on future prospects. Even if you'd like to run a live blog on some project you're working on as a sort of portfolio - do it on a platform you have full control of in case you want to rescind it or modify it later.
- There were several generalizations in that statement that align with my similar fears to the OP. Most firmware should minimize the charge cycling, most batteries should be stable at constant charge... most isn't great for something that I want to sit in the corner undisturbed for a decade just chugging along - I have a few old desktops I use whenever I need a stand alone server or to host something web-live for a while. They'll eventually have hardware failures, but I have a lot more confidence that when they fail it won't be dramatic or destructive - ditto with old laptops, the serviceability expectations are much higher than phones so I have yet to meet a laptop I can't pop open and just pull the battery out of to run on AC alone - in the case of a power failure the UPS can't cover I'd rather the machine just power off rather than needing to deal with the possibility of dramatic failure.
I think if you're considering re-harvesting old devices to use for hosting and get far enough down your list to get to phones then you've likely got enough constant maintenance costs in overseeing things that the additional worry of fire risk just isn't worth it.
- The better practice is, if it isn't a one-off, being introduced to the tool (perhaps by an LLM) and then just running the tool yourself with structured inputs when it is appropriate. I think the 2015 era novice coding habit of copying a blob of twenty shell scripts off of stack overflow and blindly running them in your terminal (while also not good for obvious reasons) was better than that essentially happening but you not being able to watch and potentially learn what those commands were.
- But we are still reliant on the LLM correctly interpreting the choice to pick the right skill. So "known to work" should be understood in the very limited context of "this sub-function will do what it was designed to do reliably" rather than "if the user asks to use this sub-function it will do was it was designed to do reliably".
Skills feel like a non-feature to me. It feels more valuable to connect a user to the actual tool and let them familiarize themselves with it (and not need the LLM to find it in the future) rather than having the tool embedded in the LLM platform. I will carve out a very big exception of accessibility here - I love my home device being an egg timer - it's a wonderful egg timer (when it doesn't randomly play music) and I could buy an egg timer but having a hands-free egg timer is actually quite valuable to me while cooking. So I believe there is real value in making these features accessible through the LLM over media that the feature would normally be difficult to use in.
- The really weird thing is that Big Business actually is buying supercomputer clusters to do just that. I can't really talk to the government side but a lot of businesses' early forays into AI was just slapping a chatbot on their product and hoping it'd attract a lot more business. I also think you'd be surprised how integrated really dumb chatbots are into business communication these days.
I think most smart people are looking seriously at different models to try and improve the accuracy of any existing ML uses they had in their business but the new features built post-ChatGPT tend to often just be fancied up chats.
- I will freely admit this - but I have led a localization effort on a mobile game and worked on localization for a desktop application. I am also lucky enough to travel abroad quite a bit and am quite familiar with consumer offerings.
So I would fairly limit my experience to consumer and medium-sized business uses - I have no experience with large corporate translation efforts (the largest would probably be Ubisoft or the Mouse-Ears company's gaming divisions if you consider them large) and even the small mobile game company I worked at had a budget in the millions range. It certainly hasn't been a focus of my career but I feel comfortable standing by my statement above.
- There was a recent Bezos talk about the fact that (much like the dot-com) we've overspending on infrastructure that'll bubble and implode but then we'll have all this amazing infra for companies to build off of... but the process of that overspend and implosion is essentially a massive debt erasure - a lot of people are currently propping up this market with their capital and the companies they're propping up will collapse and those obligations liquidated - and that will result in massive society-wide pain. We may end up in a better place for the next generation because of this investment - but if you're a retail investor don't expect your 401k accounts to weather that burst gracefully... and, unlike the boomers, this will largely hit Millenials and Gen Z both of whom are currently under massive financial stress.
- As someone working in a field that has used NLP for quite some time - yeah, I generally agree that those investments are worth their weight in gold... which is unfortunate because before ChatGPT came along they were viewed as niche unprofitable money-sinks. The astronomical investments we've seen lately have been in general models which can be leveraged to outperform some of our older models but had we wanted purely to improve those models there were much more efficient ways to do so.
Hopefully we can retain a lot of this value when the bubble bursts but I just haven't seen any really good success stories of converting these models into businesses. If you try and build as a middleman where you leverage a model to solve someone's problem they can always just go to the model runner and get the same results for cheaper - and the model runners seem (so far - this may change) to be unable to price model usage at a level that actually makes it sustainable.
Those older models running specialized tasks seem to be trucking along just fine for now - but I remain concerned that when the bubble bursts it's going to starve these necessary investments of capital.
- Ignorance is not a defense for breaking a law and I do hold my politicians (and hope other voters also hold their politicians) to a similar standard when it comes to passing a bill. Lawmakers exist because we've all got food to put on the table and can't handle the constant effort required to understand the intricacies and implications of legal code maintenance - it's the job they've been hired to do and if they're not doing it they shouldn't have that job anymore.
- Some good news though - having baked in sales tax being required in advertising actually aligns marketing lobbying with pushing for harmonized sales taxes which I'd generally consider a more just system. IMO adding random regressive taxes in different counties to make up budget shortfalls causing very strange market effects is a bad thing.
- I'd like to clarify that point a bit.
They're allowed to get away with it because of a dysfunctional lobbying driven government. Mils don't exist in the common knowledge and if any reasonable person looked at this they'd call it out. It is useful in accounting but a Mill has never been minted and the last half penny was minted in 1857. It has never been possible using issued physical legal tender in the US to pay a debt of $3.129
The Mill doesn't exist because of some archaic need - it's pure dysfunction and the utilization of it in gas prices is a practice that should and very easily could be made illegal.
- America is allergic to baked in taxes - you've got to keep the appearance of a deal even when there isn't one. America also embraces a lot of junk and hidden fees - ticketmaster is a great example of this.
I think consumers would love having baked in taxes and clear prices and were the government functional I'd hope that a consumer advocacy agency could enforce this - but that's simply not where we are right now.
- > UPS is charging customs brokerage fees of dozens or hundreds of dollars on top of the actual tariff payment
To reinforce this point... UPS just does this all the time. I had to have a number of personal effects[1] shipped up from the US to Canada that I requested self-declaration forms for them and never received them - UPS decided to brokerage the shipment themselves. We then spent the next three months fighting a six hundred dollar charge[2] that should have never existed.
UPS is going to defraud customers on brokerage fees regardless of the scenario - it's just what UPS does. You've got bigger problems to worry about - the impact of dropping the penny will be unnoticeable in the sea of general corruption and fraud.
1. Items that you own in one country and are shipping to Canada for personal possession are exempt from most normal tariffs.
2. To really add icing to outrage - this was more than double the original shipping price and, considering we delivered an itemization with the shipment for customs UPS could calculate their BS fee upfront and show the actual cost to the customer but they don't because the US doesn't force them to.
- > Boomers and older folks are not culturally or mentally equipped to handle it
I'm glad you mentioned this because the "Grandma - I was arrested and you need to send bail" scams are already ridiculously effective to run. Better TTS will make voice communication without some additional verification completely untrustworthy.
But, also, I don't want better TTS. I can understand the words current robotic TTS is saying so it's doing the job it needs to do. Right now there are useful ways to use TTS that provide real value to society - better TTS would just enable better cloaking of TTS and allow actors to more effectively waste human time. I would be perfectly happy if TTS remained at the level it is today.
- MS lives by corporate contracts and there are a lot of very powerful US companies that will roll over if Trump barks - if MS had already fled the US in a legal sense they'd definitely be in a better place but trying to leave during this administration would cause Trump's ire to focus on them and likely cost them an immense amount of money. I don't particularly like MS and both office and windows are declining in quality quickly so I wouldn't be opposed to the move but... nothing would sink that ship faster than losing a bunch of large US contracts as Trump toadies demonstrate their loyalty by bravely switching to alternatives.